Declining Test Scores in the United States Signal the Need for AI Solutions
Results from this year’s National Assessment of Education Progress (NAEP) from the U.S. Department of Education show a continued, steep decline in test scores in all subjects. Math scores for eighth graders were the lowest in 50 years, and reading scores regressed back to scores last seen in 1998. Colloquially known as the “Nation’s Report Card,” the NAEP provides national, state, and even some district-level measurements for student achievement in both public and private schools across a variety of demographics. The results are worrying and signal a lasting impact of the pandemic on learning outcomes. And students need immediate support. That’s where AI can help.
Students differ in their respective needs, skills, and learning styles. But schools in the United States frequently “teach to the middle,” with lesson plans suited for the average student in the class. While this pedagogical approach can work in some instances, when there are clear deficiencies for students in some areas, a personalized approach has proven to be effective in increasing scores. AI-enabled educational tools can support this type of individualized support while incorporating learning objectives, instructional approaches, and specific content. For example, Carnegie Learning’s AI-powered platform uses analytics to adjust the difficulty level of lessons based on a student’s individual needs, helping balance strengths and weaknesses. Students who used the company’s MATHia platform, particularly lower-performing students, performed better in their Algebra 1 classes after using the technology. Similarly, the popular non-profit educational platform Khan Academy has begun pilot testing an AI-enabled tutor that uses GPT-4 to help students work through math problems.
Generative AI technology in the classroom also opens up new opportunities for students to learn in their preferred modality. Multimodal learning helps students process information in different ways according to their preferences, which can keep them more engaged and motivated. For example, with Nolej AI, teachers can create a swath of interactive material based on an uploaded lecture document, whether it be flashcards or videos, or a gamified lesson. Students can then access the material in their preferred modality. With some AI tools for younger students, the content can even change depending on what things interest the students, like reading comprehension questions all about dinosaurs for one student and about sports for another. Keeping students engaged in the learning process will help them become stronger students in the long run.
And still, some groups hesitate to take this step. The Department of Education’s own Office of Education Technology released its long-awaited recommendations on AI in the classroom, ultimately supporting the precautionary principle, which is the idea of holding back technological deployment until all potential risks can be properly addressed. But with the NAEP results, it seems clear that the risks and harms associated with the status quo are equally, if not more, undesirable than the potential risk from AI.
Students are not learning or absorbing the skills and content they need. And in the face of a national teachers shortage, shouldn’t they receive every possible solution? The Department of Education should use these results as a wake-up call that AI should not be kept on the sidelines, especially if it can make a tangible difference in student outcomes. The Department should sponsor grant programs to accelerate the use of AI in the classroom, particularly in lower-resource schools.
While AI certainly needs careful scrutiny in some areas, in the case where it can clearly help children learn, its use should be a no-brainer.
